Search

Finding the optimum material with improved properties for a given application is challenging because data acquisition in materials science and chemistry is time consuming and expensive. Therefore, dealing with small datasets is a reality in chemistry, whether the data are obtained from synthesis or computational experiments. In this work, we propose a new artificial intelligence method based on active learning (AL) to guide new experiments with as little data as possible, for optimum experimental design. The AL method is applied to ABO3 perovskites, where a descriptor based on atomic properties was developed. Several regressor algorithms were employed: artificial neural network, Gaussian process, and support vector regressor. The developed AL method was applied in the experimental design of two important materials: non-stoichiometric perovskites (Ba(1− x)A x Ti(1− y)B y O3) due to substituting ionic sites with different concentrations and elements (A = Ca, Sr, Cd; B = Zr, Sn, Hf), aiming at the maximization of the energy storage density, and stoichiometric ABO3 perovskites where different elements are changed in the A and B sites for the minimization of the formation energy. AL for experimental design is implemented in the machine learning agent for chemistry and design (MLChem4D) software, which has the potential to be applied in inorganic and organic synthesis (e.g., search for the optimum concentrations, catalysts, reactants, temperatures, and pH to improve the yield) and materials science (e.g., search the periodic table for the proper elements and their concentrations to improve the materials properties). The latter marks the first MLChem4D application for the design of perovskites.

Euterpe edulisMartius is an endangered species of the Atlantic Forest, whose fruits have high antioxidant potential, and propagated exclusively by seeds. The present study assessed the ability of different auxin inducers and picloram analogs to trigger somatic embryogenesis inE. edulis. Immature seeds were harvested, and their zygotic embryos were excised and grown in MS culture medium supplemented with 2,4-D dichlorophenoxyacetic acid (2,4-D) or picloram at 150, 300, 450, 600 µM. The activity of picloram analogs triclopyr and clopyralid was evaluated in semisolid MS medium. At maturation and germination, picloram-derived calli and somatic embryos isolated from triclopyr-grown cultures were first transferred to pre-maturation medium and, after 30 days, to basal MS or MS medium supplemented with either 5 µM abscisic acid or 0.53 µM 1-naphthaleneacetic acid plus 12.3 µM 2-isopentenyladenine. Finally, somatic embryos with root protrusions were transferred to MS medium devoid of sucrose for 30 days and then acclimatized ex vitro. Scanning, transmission, and atomic force microscopy revealed that picloram was superior to 2,4-D but less effective than triclopyr (100 µM) in inducing embryogenesis. Maturation and germination of somatic embryos in E. edulis can be maximized by 5 µM abscisic acid, and selecting calli via atomic force microscopy.HighlightThis work opens novel roads for embryogenic induction, using a new and more efficient inducer than the usual ones, and an innovative evaluation technique based on AFM.

Strawberry is a food rich in bioactive compounds with great antioxidant potential. However, due to the high incidence of pests that affect crop cultivation, phytosanitary management still lacks control methods for agroecological cultivation. Thus, the present research aimed to evaluate the chemical composition and the potential of the essential oil of the leaves of Piper macedoi in the control of Cerosipha forbesi in laboratory and semi-field conditions. The concentration of essential oil in the leaves of P. macedoi that showed the highest mortality was 2.0 ml/L of oil, with a mortality above 91% under laboratory conditions. A mortality rate of 80% for all concentrations tested was observed after 24 h in all conditions tested. Thus, using essential oil from the leaf of P. macedoi can be a highly viable strategy in managing the aphid C. forbesi since it showed high mortality rates with small doses of oil.

Adsorbate interactions with substrates (e.g. surfaces and nanoparticles) are fundamental for several technologies, such as functional materials, supramolecular chemistry, and solvent interactions. However, modeling these kinds of systems in silico, such as finding the optimum adsorption geometry and energy, is challenging, due to the huge number of possibilities of assembling the adsorbate on the surface. In the current work, we have developed an artificial intelligence (AI) approach based on an active learning (AL) method for adsorption optimization on the surface of materials. AL uses machine learning (ML) regression algorithms and their uncertainties to make a decision (based on a policy) for the next unexplored structures to be computed, increasing, though, the probability of finding the global minimum with a small number of calculations. The methodology allows an accurate and automated structural elucidation of the adsorbate on the surface, based on the minimization of the total electronic energy. The new AL method for adsorption optimization was developed and implemented in the quantum machine learning software/agent for material design and discovery (QMLMaterial) program and was applied for C60@TiO2 anatase (101). It marks another software extension with a new feature in addition to the automatic structural elucidation of defects in materials and of nanoparticles as well. SCC-DFTB calculations were used to build the complex search surfaces with a reasonably low computational cost. An artificial neural network (NN) was employed in the AL framework evaluated together with two uncertainty quantification methods: K-fold cross-validation and non-parametric bootstrap (BS) resampling. Also, two different acquisition functions for decision-making were used: expected improvement (EI) and the lower confidence bound (LCB).

The SCC-DFTB repulsion parameters based on the material science set (matsci) were redesigned to describe the structure and dynamic properties of bulk liquid water. The iterative Boltzman inversion ...

Active learning (AL) has been successfully applied in materials science for the global optimization of clusters and defects in materials. Many important chemistry problems require the structural elucidation of molecules as a first step to the mechanistic elucidation of complex heterogeneous catalysis phenomena. Theoretical methods coupled with global optimization algorithms are successfully used for this purpose. However, it is challenging to find the global minimum structure together with the proper electronic spin multiplicity. In this work, we present an AL implementation for global optimization of atomic clusters where the spin multiplicity is considered in the search loop (SM@AL). The method was implemented in the QMLMaterial software, interfaced with the deMon2k program to perform local structure optimizations. In this work, we present applications of SM@AL for the global optimization, in terms of molecular structure and electronic spin of 3Al@Si11, where Si11 is doped by 3 Al, and Mo4C2 with spin multiplicities 2, 4 and 6 and 1, 3, 5, 7, 9 and 11, respectively., 20th deMon Developers Workshop

In catalysis, an accurate structural elucidation of molecules, atomic clusters, nanoparticles and solid surfaces is required to understand chemical processes. Therefore, an efficient and automatic structure determination for these systems is of great benefit since it requires a global search within huge chemical spaces. In this work, we propose a new active learning (AL) method intended for atomic clusters that uses different supervised machine learning techniques and their uncertainties to decide the promising non-observed (virtual) structures to be evaluated from quantum calculations. The method was developed for structural elucidation of (I) clusters where all atomic coordinates are allowed to change in a continuous chemical space and (II) doped ones where the atoms exchange in a sufficiently rigid structure, thus, a discrete search space where all cluster descriptors are known. Particularly for case I, a genetic algorithm operator was used to create the unknown virtual structures (and then their descriptors) from the observed ones to improve the performance of the AL search. The proposed AL was applied to the global optimization of heteronuclear (Al4Si7 and 4Al@Si11) and homonuclear (Na20) clusters using self-consistent charge density-functional tight-binding (SCC-DFTB), where a new repulsion parameter was developed to reproduce isomers evaluated from high-level calculations. The performance of the Gaussian process and artificial neural network algorithms was evaluated together with several uncertainty quantification methods: from Gaussian process, K-fold cross-validation and nonparametric bootstrap (BS) resampling. The efficiency of the AL was compared to conventional global optimization methods, such as random search and a genetic algorithm. The results show that the AL can find efficiently the global minimum of atomic clusters.

The study of materials in the nanoscale regime has important applications for catalytic reactions, the energy industry and medicine. We performed exploratory density functional theory calculations for molybdenum disulphide (MoS2) and calcium carbonate (CaCO3) nanoparticles (NPs), the former being developed as a hydrogenation and coke-prevention catalyst and the latter representing the catalytic support in some reservoirs. Utilizing Born–Oppenheimer Molecular Dynamics with reduced masses as a global optimization method, we found Mo8S16 and Mo16S32 NPs with lower-lying energies than those of locally optimized crystal geometries. Our results suggest that MoS2 NPs prefer a tetragonal lattice arrangement which is in agreement with previous studies of smaller MoS2 NPs. It remains to be seen how prenucleation MoS2 clusters of the hexagonal phase are formed. The CaCO3 NPs showed small energy differences between vastly different conformations. Therefore, in order to capture only their supporting effects in reservoirs, one should consider keeping a substantial part of the models fixed.

Designing and understanding the mechanism of non-stoichiometric materials with enhanced properties is challenging, both experimentally and even computationally, due to the large number of chemical spaces and their distributions through the material. In the current work, it is proposed a Machine Learning approach coupled with the Efficient Global Optimization (EGO) method—an Adaptive Design (AD)—to model local defects in materials from first-principle calculations. Our method takes into account the smallest sample set as possible, envisioning the material defect structure relationship with target properties for new insights. As an example, the AD framework allows us to study the stability and the structure of the modified goethite (Fe0.875Al0.125OOH) by considering a proper defect distribution, from first-principle calculations. The chemical space search for the modified goethite was evaluated by starting from different sizes and configurations of the samples as well as different surrogate models (ANN and Gaussian Process; GP), acquisition functions, and descriptors. Our results show that the same local solution of several defect arrangements in Fe0.875Al0.125OOH is found regardless of the initial sample and regression model. This indicates the efficiency of our search method. We also discuss the role of the descriptors in the accelerated global search for defects in material modeling. We conclude that the AD method applied in material defects is a successful approach in automating the search within huge chemical spaces from first-principle calculations by considering small samples. This method can be applied to mechanistic elucidation of non-stoichiometric materials, solid solutions, alloys, and Schottky and Frenkel defects, essential for material design and discovery.

In this work, we explore the possibility of using computationally inexpensive electronic structure methods, such as semiempirical and DFTB calculations, for the search of the global minimum (GM) structure of chemical systems. The basic prerequisite that these inexpensive methods will need to fulfill is that their lowest energy structures can be used as starting point for a subsequent local optimization at a benchmark level that will yield its GM. If this is possible, one could bypass the global optimization at the expensive method, which is currently impossible except for very small molecules. Specifically, we test our methods with clusters of second row elements including systems of several bonding types, such as alkali, metal, and covalent clusters. The results reveal that the DFTB3 method yields reasonable results and is a potential candidate for this type of applications. Even though the DFTB2 approach using standard parameters is proven to yield poor results, we show that a re-parametrization of only its repulsive part is enough to achieve excellent results, even when applied to larger systems outside the training set.

Oxide-derived copper (OD-Cu) catalysts are promising candidates for the electrochemical CO2 reduction reaction (CO2RR) due to the enhanced selectivity toward ethylene over methane evolution, which has been linked to the presence of subsurface oxygen (Osb). In this work, Osb is investigated with theoretical methods. Although Osb is unstable in slab models, it becomes stabilized within a “manually” reduced OD-Cu nanocube model which was calculated by self-consistent charge density functional tight binding (SCC-DFTB). The results obtained with SCC-DFTB for the full nanocube were confirmed with subcluster models extracted from the nanocube, calculated with both density functional theory (DFT) and SCC-DFTB. The higher stability of Osb in the nanocube is attributed to the disordered structure and greater flexibility. The adsorption strength of CO on Cu(100) is enhanced by Osb withdrawing electron density from the Cu atom, resulting in reduction of the σ-repulsion. Hence, the coverage of CO may be increased, fac...

Pyrite is the most common sulfide in the Earth. In the presence of arsenopyrite its oxidation is delayed, and instead, the arsenopyrite increases its oxidation rate, releasing As(III) and As(V) species in the medium. DFT/plane waves calculations were performed on pyrite/arsenopyrite interface models to understand the stability, structure, and electronic properties of the interface. This is the first step to understand the influence of the inlaid arsenopyrite in the pyrite oxidation mechanism. The interface is slightly stressed with minor changes in the bond lengths and lattice parameters with respect to the pure phases. The work of adhesion and the formation energy indicate that the miscibility of the two phases is not favorable, explaining the presence of large domains of either pyrite or arsenopyrite forming bulk granular regions. The valence band of the pyrite/arsenopyrite interface has large contributions from the pyrite phase, while the conduction band has large contributions from the arsenopyrite. T...

The aim of this work was to study the interaction between the local anesthetic benzocaine and p-sulfonic acid calix[n]arenes using NMR and theoretical calculations and to assess the effects of complexation on cytotoxicity of benzocaine. The architectures of the complexes were proposed according to (1) H NMR data (Job plot, binding constants, and ROESY) indicating details on the insertion of benzocaine in the cavity of the calix[n]arenes. The proposed inclusion compounds were optimized using the PM3 semiempirical method, and the electronic plus nuclear repulsion energy contributions were performed at the DFT level using the PBE exchange/correlation functional and the 6-311G(d) basis set. The remarkable agreement between experimental and theoretical approaches adds support to their use in the structural characterization of the inclusion complexes. In vitro cytotoxic tests showed that complexation intensifies the intrinsic toxicity of benzocaine, possibly by increasing the water solubility of the anesthetic and favoring its partitioning inside of biomembranes.

We have investigated the structure and electronic structure of single- and double-walled imogolite nanotubes with Ge and Si as group IV element. While it is known from experiment, and in the case of single-walled tubes confirmed by theory, imogolite nanotubes are monodisperse in diameter. We show that imogolite tubes are also showing a preferred chirality (zigzag), resulting from the hydrogen-bond network on the tube surfaces, and that there is an exceptionally stable form of intertube interaction that supports the formation of monodisperse double-walled imogolite nanotubes. The strongest stabilization of double-walled tubes has been found for tube indexes with nine units of difference around the circumference, and the minimum structure is found for the (12,0)@(21,0) tube in the case of germanium imogolite and (9,0)@(18,0) for imogolite. The electronic structure is only slightly affected by these geometric factors, as are the mechanical properties, which show Young moduli of 320–370 GPa, thus being in the...

The self-consistent-charge density-functional tight-binding method (SCC-DFTB) is largely used for investigating systems of increasing complexity. However, the SCC-DFTB parameters are not available for the whole periodic table and some issues related to its transferability can limit its usefulness. The framework for automating the repulsion energy (E rep) parameterization uses reference chemical models to parameterize the repulsion term of the SCC-DFTB. The aim is to provide a straightforward and accessible manner to obtain accurate SCC-DFTB E rep parameters automatically. Large number of reference chemical models can be simultaneously used to increase the accuracy and transferability of the parameters with respect to the total energies and derived properties. SCC-DFTB parameters for the C, O, N, H set were obtained. The results are in good agreement with the well-established SCC-DFTB parameters available for this test case. Extensions to include other properties such as reaction energies in the E rep parameterization are briefly discussed.

Structural, electronic, and mechanical properties of single-walled chrysotile nanotubes have been investigated using the self-consistent charge density-functional tight-binding method (SCC-DFTB). T...

The electronic, structural and mechanical properties of the modified imogolites have been investigated using self consistent charge-density functional-tight binding method with “a posteriori” treatment of the dispersion interaction (SCC-DFTB-D). The zigzag (12,0) imogolite has been used as the initial structure for the calculations. The functionalization of the interior (12,0) imogolite nanotubes by organosilanes and by heat treatment leading to the dehydroxylation of the silanols were investigated. The reaction of the silanols with the trimethylmethoxysilanes is favored and the arrangement of the different substitutions that leads to the most symmetrical structures are preferred. The Young moduli and band gaps are slightly decreased. However, the dehydroxylation of the silanol groups in the inner surface of the imogolite leads to the increase of the Young moduli and a drastic decrease of the band gap of about 4.4 eV. It has been shown that the degree of the dehydroxylation can be controlled by heat treatment and tune the band gap, eventually, leading to a semiconductor material with well defined nanotube structure.

Imogolites (aluminosilicate and aluminogermanate) are nanotubes (NTs) with well-defined diameter and chirality that can be synthesized in aqueous solution in mild conditions. It has been shown that the diameter and the single- or double-walled NT formation can be controlled depending on the synthesis parameters such as anionic ligands, temperature, pH and reactant concentration. In this chapter, the advances for understanding at a molecular level the remarkable properties of imogolites, their stability and formation mechanism are reviewed highlighting the computational chemistry contribution. Different theoretical methods have been used, in special the Self-Consistent-Charge Density-Functional Tight-Binding method (SCC-DFTB), the most used for investigating imogolites, is here briefly described.

Imogolite is a single-walled aluminosilicate nanotube (NT) found in nature that can be easily synthesized, as well as its analogue aluminogermanate NT. Based on geometrical assumptions and pKa values, species such as H3PO4, H3PO3, H3AsO3, H3AsO4 could also be candidates to form imogolite-like structures. In the present work, we provide insights about the stability, electronic, structural and mechanical properties of possible imogolite like NTs by means of self-consistent charge density-functional tight-binding method (SCC-DFTB). Similarly to aluminogermanate, where the tetrahedral silicate groups are replaced by germanate, in this work tetrahedral silicate groups are substituted by phosphate, phosphite, arsenate and arsenite units in the imogolite structure. Detailed analysis is focused on structural properties, strain energy, band gap and Mulliken charges distribution. The calculated strain energy curves for all studied zigzag imogolite-like NTs present well-defined minima, which change as a consequence of composition variation. Moreover, the strain energy curves of armchair imogolite-like NTs also present minima, although in all cases less stable than zigzags by at least 2.2 meV per atom. The insulating NT behaviour changes after internal modification from silicate to phosphate, phosphite, arsenate and arsenite, as well as the charge distribution inside and outside the nanotubes.

Self-consistent-charge density-functional tight-binding (SCC-DFTB) approximated method was employed to investigate the structural, mechanical and electronic properties of the zigzag and armchair nano-fibriform silica (SNTs) and their outer surface organic modified derivatives (MSNTs) with internal radii in the range of 8 to 36 Å. The strain energy curves showed that the nanotubes structures are energetically more stable compared to the respective sheet structures. External hydroxyl dihedral angles in silica nanotubes have small influence, about 0.5 meV.atom(-1), in the strain energy curve tendency of those materials favoring the zigzag chirality. The chemical modification of outer surface of SNTs by dimethyl silane group affects their relative stability favoring the armchair chirality in approximately 2 meV.atom(-1). MSNTs have axial elastic constants, Young's moduli, determined at the harmonic approximation, around 100 GPa smaller than the respective SNTs. The Young's moduli of zigzag and armchair SNTs are in the range of 150-195 GPa and 232-260 GPa, respectively. And for the zigzag and armchair MSNTs these values are in the range of 77-89 and 110-140 GPa, respectively. The SNTs and MSNTs were characterized as insulators with band gaps around 8-10 eV.

The emerging field of nanotechnology is mostly focused on carbon and inorganic based nanomaterials, such as carbon nanotubes, graphene, transition metal nanotubes and nanowires (Iijima, 1991; Tenne et al., 1992; Endo et al., 1996; Dresselhaus et al., 2001). Systems containing aluminosilicates have been investigated as mesoporous materials in the form of zeolite and alumina. Although they have not yet received as much attention, clay minerals can also form nanostructured layered materials and nanotubes with remarkable geometric properties. Imogolite is the most representative species of this case, since it has been studied in a pre-nano (1970) decade (Cradwick et al., 1972) and has been nearly forgotten until recently. Since 2000 (Bursill et al., 2000; Tamura & Kawamura, 2002; Mukherjee et al., 2005; Nakagaki & Wypych, 2007), these structures gained again prominence in the literature and appear as an emerging field of research. They can be used as nanoreactors for selective catalysts, adsorbent, nanocable, support for the immobilization of metalloporphyrins, encapsulation and ionic conductor (Nakagaki & Wypych, 2007; Kuc & Heine, 2009).

Em escala nanométrica as propriedades de agregados atômicos e moleculares são altamente dependentes de seus tamanhos e se alteram de modo não uniforme com o crescimento do sistema. Dado que essas propriedades são distintas daquelas do sistema macroscópico, novos tipos de materiais com diferentes aplicações podem ser obtidos. Atenção específica é dada para clusters metálicos, frequentemente aplicados em catálise, e para clusters atômicos de nitrogênio, que têm chamado a atenção da comunidade científica por serem bons candidatos a materiais de alta densidade de energia. Existe um interesse crescente atualmente no emprego de materiais nanométricos como novos possíveis tipos de combustíveis. Para ser considerado eficiente, um material combustível deve ser suficientemente instável de modo a reagir com outro sistema e liberar quantidade significativa de energia. Por outro lado, deve ser estável o suficiente para ser sintetizado e estocado sem se decompor espontaneamente. Ademais, esses materiais devem gerar produtos consideravelmente menos tóxicos que aqueles já difundidos hoje em dia, além de apresentarem melhores performances e serem técnica e economicamente viáveis. Nesse contexto, sistemas de polinitrogênio têm sido estudados devido à sua dissociação em moléculas de N2 inertes, processo que pode liberar enormes quantidades de energia. A abordagem comum no estudo de clusters envolve o uso inicial de potenciais analíticos empíricos como um modo rápido de se avaliar a energia do sistema, explorando-o através de métodos de otimização global como os algoritmos genéticos. Tais potenciais muitas vezes oferecem uma descrição pobre do sistema, especialmente no regime de poucos átomos, já que são incapazes de reproduzir efeitos quânticos. Portanto, um algoritmo genético acoplado a cálculos de estrutura eletrônica foi desenvolvido para, em um primeiro momento, buscar pelo mínimo global na superfície de energia potencial ab initio de nanoligas. Isso permite uma busca imparcial sobre novas estruturas que não poderiam ser reveladas por potenciais empíricos. O critério de convergência para o gradiente da energia nas otimizações locais é progressivamente reduzido ao longo das gerações, diminuindo o número de avaliações de energia necessárias. Um caso estudado foi a liga Na-K em nível MP2/ECP, onde a eficiência do algoritmo na obtenção do mínimo global foi demonstrada. Uma análise particular da transição 2D-3D-2D para o caso de 6 átomos foi apresentada em detalhes pela primeira vez. Estruturas completamente novas também foram obtidas para clusters maiores. Em sequência, esse algoritmo genético quântico foi aprimorado e aplicado na exploração de hipersuperfícies de energia potencial DFT de clusters atômicos de nitrogênio de camada fechada de até 10 átomos. Um parâmetro de corte para a distância interatômica, definido pelo usuário, e um ¿gerenciamento de operadores¿ foram adicionados ao procedimento evolucionário padrão. Os funcionais de troca e correlação B3LYP e PBE foram explorados com as funções de base 6-31G usando o algoritmo. Posterior reavaliação das estruturas geradas foi realizada pela reotimização e análise vibracional em níveis MP2 e CCSD(T) empregando funções de base maiores e consistentes com correlação. As energias de ligação de todas as estruturas estáveis de polinitrogênio encontradas foram calculadas e comparadas, assim como suas energias de dissociação em moléculas de N2, N3+ e N3-. Com a presente abordagem, estruturas de polinitrogênio previamente reportadas foram confirmadas, assim como a estabilidade de novas estruturas puderam ser previstas. Pode-se também concluir que o perfil da superfície de energia potencial depende claramente do método de cálculo empregado. Investigações teóricas das curvas de energia potencial eletrônica associadas à dissociação do N4 (Td) em 2N2 também foram realizadas para o estado fundamental e primeiros estados excitados utilizando CASSCF(12,12). Os gráficos contendo os caminhos de dissociação envolveram estados singleto e tripleto, onde foi possível observar interseções cônicas e cruzamentos intersistema. Possíveis canais de dissociação alternativos foram inferidos a partir desses cruzamentos e da fotoexcitação dessa espécie. Evidências foram obtidas para resolução de um possível mecanismo de reação envolvendo a abstração de um átomo de N de uma molécula de N2 em um estado eletrônico excitado, por outra molécula de N2 excitada, para produzir N3 linear. At the nanoscale the properties of an atomic or molecular aggregate depend largely on its size, changing in a non-uniform pattern as the system increases. Given that these properties are distinct from the bulk matter, new types of materials with different applicability may be found. Specific focus is given to metallic clusters, often with potential application in catalysis, and to atomic nitrogen clusters, which have been lately calling the attention of the scientific community for being goodcandidates for high energy density materials. There is a rising interest nowadays in employing nanoscale materials as new possible types of fuels. To be considered an efficient fuel, the candidate material must be sufficiently unstable in order to react with another system and then release a considerable amount of energy. On the other hand, it must be stable enough to be synthesized and stored without spontaneously decomposing itself. To accomplish the goal of being considered greenthese materials must not only generate reasonably less toxic products than the current widely spread fuels do, but also demonstrate higher performances while being technically as well as economically attractive. In this context, polynitrogen systems have been studied due to their dissociation into inert N2 molecules that may release huge amounts of energy. The normal approach taken in the study of clusters is to first use an empirical analytic potential, such as to have a fast way of evaluating the energy, and explore it using global optimization methods such as genetic algorithms. Such empirical potentials often turn out to be a poor representation of the system, especially in the few atoms regime,since they cannot reproduce quantum effects. Therefore, a genetic algorithm coupled to electronic structure calculations was developed for searching, in the first instance, the global minimum in the ab initio potential energy landscape of alloy clusters. This allows a non-biased search over new structures that could not be unraveled by empirical potentials. The convergence criterion for the energy gradient in local optimization is progressively reduced over the generations in order to require less energy evaluations. A case studied is the Na-K alloy with MP2/ECP energies, where it is demonstrated that the algorithm is efficient in obtaining global minima. Particular analysis of a 2D-3D-2D transition in the 6-atoms case is studied in detail for the first time. Completely new structures are unveiled for larger alloy clusters. Subsequently, this quantum genetic algorithm was upgraded and employed to survey the DFT potential energy hypersurfaces of closed-shell atomic nitrogen clusters up to ten atoms. An atom-atom distance threshold parameter, controlled by the user, and an operator manager were added to the standard evolutionary procedure. Both B3LYP and PBE exchange-correlation functionals with 6-31G basis set were explored using the algorithm. Further evaluation of the structures generated was performed through reoptimization and vibrational analysis within MP2 and CCSD(T) levels employing larger correlation consistent basis set. The binding energies of all stable polynitrogen structures found were calculated and compared, as well as their energies relative to the dissociation into N2, N3+ and N3- molecules. With the present approach we confirmed some previously reported polynitrogen structures and predicted the stability of new ones. We can also conclude that theenergy surface profile clearly depends on the calculation method employed. Theoretical investigations of the electronic potential energy curves associated with the dissociation of the N4 (Td) into two N2molecules were also performed for the electronic ground state and the first excited states using CASSCF(12,12) level of theory. The graphs containing the dissociation paths studied involved singlet and triplet states, where it was possible to observe both conical intersections and intersystem crossings. Possible alternative dissociation channels for the N4 (Td) were then inferred from these crossings and photoexcitation of this species. Insights were also provided to unravel a possible reaction mechanism involving the abstraction of a nitrogen atom from a N2 molecule in an excited electronic state, by another excited N2 molecule, to produce linear N3.